A Reinforcement Learning Based Dynamic Walking Control

A quasi-passive dynamic walking robot is built to study natural and energy-efficient biped walking. The robot is actuated by MACCEPA actuators. A reinforcement learning based control method is proposed to enhance the robustness and stability of the robot's walking. The proposed method first learns the desired gait for the robot's walking on a flat floor. Then a fuzzy advantage learning method is used to control it to walk on uneven floor. The effectiveness of the method is verified by simulation results.

Authors

• Yong Mao
• Jiaxin Wang
• Peifa Jia
• Shi Li 0002
• Zhen Qiu
• Le Zhang
• Zhuo Han

Keywords

• Learning
• Legged locomotion
• Robust stability
• Knee
• Robot sensing systems
• Robotics and automation
• Robot control
• Leg
• Actuators
• Energy efficiency
• Dynamic Walking
• Actuator
• Natural Walking
• Flat Floor
• Collision
• Learning Algorithms
• Local Actors
• Phase Change
• Knee Joint
• Lagrange Multiplier
• Hip Joint
• Joint Angles
• Digital Signal Processing
• Learning Control
• Fuzzy Rules
• Joint Velocity
• Learning Targets
• Rough Terrain
• Walk Process
• Walking Gait
• Joint Trajectories
• Action-value Function
• Uneven Terrain
• Value Of Rule
• Degrees Of Freedom
• Patellar
• Dynamic Model
• Membership Function
• passive dynamic walking
• reinforcement learning
• fuzzy advantage learning
• biped robot